Vector Databases

Vector databases store and search over high-dimensional numeric embeddings using similarity metrics like cosine distance, enabling semantic search that finds documents by meaning rather than exact keyword match, for example letting an agent query a codebase with "functions that handle authentication" instead of a grep-style string search. They form the storage layer for most retrieval-augmented generation (RAG) systems, providing the infrastructure that gives agents access to knowledge that does not fit within a single context window. Key options include Pinecone, Weaviate, Chroma, Qdrant, and pgvector for Postgres, each with different trade-offs around managed versus self-hosted deployment, query performance, and scale.